Emissions regulations are continually being tightened to improve air quality in many locations around the world. Over the past 30 years, the regulations relating to spark ignited engines have been tightened and the allowable emissions substantially reduced. These engines operate at or near a stoichiometric air/fuel ratio and as a result three-way catalyst technology has been developed to control the emissions of carbon monoxide (CO), unburned hydrocarbons (UHC), and nitrogen oxides (NOX), including both NO and NO2. Three-way catalyst technology is not applicable to lean burn engines because the large excess of oxygen in the exhaust mixture impedes the reduction of NOX. This is particularly a problem with diesel engines or compression ignition engines which have very high emissions of NOX and particulate matter (PM). Coupled with this is the drive within the U.S. and much of the world for increased fuel efficiency. Diesel engines are very efficient and therefore a desirable power source for vehicles. However, the high emissions must be reduced to comply with statutory requirements. To reach the emissions levels required for gasoline spark ignited engines, the emissions from a modem diesel engine must be reduced by a factor of 10 to 50, depending on the engine.
Lean-burn engines include both spark-ignition (SI) and compression-ignition (CI) engines. Lean-burn SI engines offer 20–25% greater fuel economy and CI engines offer 50% and sometimes higher fuel economy than equivalent conventional SI engines. CI engines are widely used in heavy-duty vehicles, and their use in light-duty vehicles is small but expected to grow. They are also widely used in stationary applications, such as electric power generators.
Current automotive emission control technology is based largely on catalytic converters with three-way catalysts (TWCs). This technology is highly effective for ordinary gasoline engines that operate at nearly stoichiometric air/fuel ratios. However, as discussed above, it is incompatible with lean-burn engines due to the excess of oxygen in the exhaust. This incompatibility is a major limitation of both lean-burn engines and TWC-based emission control technology. In the case of diesel engines, the emission control system must remove NOX and PM from an exhaust stream containing about 6–15% oxygen.
Many different technologies have been investigated for NOX removal from lean-burn engine exhaust. One successful technology has been the selective reduction of NOX with ammonia (NH3) as a reducing agent. Ammonia is added to the exhaust stream in an amount proportionate to the amount of NOX. The exhaust stream containing NOX and NH3 is then passed over a catalyst upon which the NOX and NH3 selectively react to produce N2. This technology, referred to as Selective Catalytic Reduction (SCR), is widely used in gas turbines and large boilers and furnaces, and is capable of achieving very high NOX conversion to N2. However, one disadvantage of this technology is that it requires a source of NH3 which can be either liquid NH3 stored under high pressure or an aqueous solution of urea which decomposes prior to or on the SCR catalyst to produce ammonia. In general NH3-SCR technology is limited to large stationary applications, due to its cost and the need for a source of NH3. In addition, the addition of NH3 must be carefully controlled to achieve a desired NH3/NOX ratio, to prevent excess NH3 from being exhausted to the atmosphere and adding to the level of air pollutants. Another disadvantage is the need to develop the rather costly infrastructure to supply ammonia or urea to vehicles using this technology. For these reasons, this technology is not the preferred approach to NOX control on vehicles or very small systems that may be located in populated areas.
Another technology that has been explored for NOX abatement from lean-burning mobile sources is a NOX storage and reduction (NSR) system, as described in Society of Automotive Engineers papers SAE-950809 and SAE-962051, and in U.S. Pat. No. 6,161,378. The NSR system has an adsorbent-catalyst unit situated in the exhaust system and through which the exhaust stream flows. This catalyst unit provides two functions: reversible NOX storage or trapping, and NOX reduction. During normal engine operation, while the exhaust gas flows through the system, NOX is adsorbed onto the adsorbent in the presence of excess oxygen during the Adsorption Cycle. As the adsorbent component becomes saturated with NOX, the adsorption becomes less complete and the NOX exiting the NOX trap begins to increase. At this point, the composition of the exhaust stream is changed from an oxidizing to a reducing state. This requires reduction of the oxygen level to zero and introduction of a reducing agent. In the reducing environment, the NOX is desorbed from the adsorbent and then reduced to nitrogen by the catalytic components that are incorporated into the adsorbent-catalyst unit. This reaction is generally very quick. Thus, the reduction part of the cycle can be very short, but must be sufficiently long to regenerate a significant fraction of the NOX adsorption capacity. The exhaust composition is then reverted to normal oxidizing conditions, and the cycle is repeated. There are several disadvantages to this technology. One problem is that converting the exhaust to reducing conditions is difficult to achieve for a lean-burn engine such as a diesel engine designed to run with 8 to 15% O2 in the exhaust stream. Another problem is that the adsorbent-catalyst components that have been investigated form very stable sulfates, resulting in poisoning of the catalyst by sulfur in the fuel. Regeneration of the catalyst to remove the sulfur is very difficult and results in degradation of the catalyst performance.
A promising approach to NOX removal from exhaust streams containing excess O2 is selective catalytic reduction of the NOX with a reducing agent such as CO or an added hydrocarbon, using a catalyst called a selective lean NOX catalyst (“lean NOX catalyst”) Such catalysts have been extensively investigated over the last 20 years (see, for example, Shelef (1995) Chem. Rev. 95:209–225, and U.S. Pat. No. 5,968,463). Previously, hydrocarbons have been used as the reducing agent, with the rationale that this component would be available from the engine fuel. In general, in engine tests using a lean NOX catalyst, when a reactive hydrocarbon is used as the sole reducing agent or fuel is injected into a diesel engine in such a manner as to produce reactive hydrocarbon species, the level of NOX control is low, in the range of 20–50%.
Hydrogen has also been found to be a good reducing agent for the selective reduction of NOX to N2. For example, Costa, et al. (2001) J. Catalysis 197:350–364, report high activity of H2 as a reducing agent for catalytic reduction of NOX in the presence of excess O2 at low temperatures (150–250° C.), with good utilization of the H2. EP 1,094,206A2 also describes beneficial results associated with addition of H2 to a hydrocarbon reducing agent in a lean NOX catalyst system, resulting in greater than 95% NOX removal in engine dynometer testing. H2/CO mixtures have also been found to be good reducing agents in such systems.
Although H2 and H2/CO mixtures are good reducing agents for continuous removal of NOX from an O2-containing exhaust stream, present methods for delivering these reducing agents for use in a small mobile system such as a vehicle are cumbersome and/or expensive. Hydrogen is difficult to store and H2 refueling stations are currently not available. On-board manufacturing of H2 or H2/CO mixtures from diesel fuel is possible, but difficult and costly.
An example of a system for on-board generation of a reducing agent for NOX reduction may be found in WO 01/34950, which describes a partial oxidation system that uses air and the on-board hydrocarbon fuel to generate a reducing mixture that is added to the exhaust stream. The exhaust stream, which contains NOX, and the added reducing agent are then reacted over a catalyst that reduces the NOX in the presence of excess O2. A disadvantage with such a device is that it may be difficult to operate for an extended period of time due to formation of coke, which ultimately poisons the catalyst. Also, this system produces low molecular weight hydrocarbons, which are less effective reducing agents for NOX than H2. Another system has been described in U.S. Pat. No. 6,176,078 that involves use of a plasmatron to produce low molecular weight hydrocarbons and H2 from hydrocarbon fuel. Disadvantages with this system include high energy cost, cost of the system including electronics for the plasma generator and durability issues. U.S. Pat. No. 5,441,401 and EP 0,537,968A1 describe use of a separate H2 generator with a separate air and water intake. Since the water is vaporized and passes through the catalyst, it must be very pure. This would require separate tanks, supply system and complexity. However, this system may be difficult to implement and too complex for NOX removal in mobile systems such as a vehicle. Another well-known technology includes an autothermal reformer (ATR) with a heat exchanger and water feed pumps. However, such a system is difficult to scale down. In addition, these processes that convert the liquid hydrocarbon fuel to H2 and CO in a separate reactor system can have a long start up time, from 1 to 30 minutes. This would result in a long period of time during which no reducing agent is available for NOx reduction and vehicle NOx emissions levels would remain unacceptably high.
Both O2 and H2O may be used to convert a hydrocarbon fuel such as diesel fuel to H2 and CO, through reactions such as partial oxidation and steam reforming in the presence of an appropriate catalyst. One approach that has previously been used for processing a hydrocarbon fuel to produce H2 and CO is to add the fuel continuously to a gas stream upstream of a catalyst, which then converts the fuel to H2 and CO when the fuel-containing gas stream contacts the catalyst. However, the disadvantage of continuous fuel addition is that the high level of O2 in the exhaust stream results in a very high temperature at a fuel-to-oxygen ratio that is good for reforming the fuel to H2 and CO. This is depicted schematically in FIG. 1. FIG. 1 depicts the reactor temperature over time at various equivalence ratios (Φ)). As fuel is added to an oxygen containing gas stream, for example a diesel exhaust containing 10% O2, combustion of the fuel results in heat release and an increase in temperature. Thus, at an equivalence ratio of 0.2, the exhaust gas would increase in temperature from about 250° C. to about 500° C. At an equivalence ratio of 0.5, the temperature would be 820° C., and at an equivalence ratio of 1.0, the temperature would be 1230° C. As the equivalence ratio rises above 1, the temperature decreases due to endothermic reforming reactions. Thus, at an equivalence ratio of 2, the temperature would be 1042° C., and at an equivalence ratio of 3, a very rich mixture, the temperature would be 845° C. Typical auto-thermal reformers, which regulate temperature isothermally by using combustion (an exothermic reaction) to supply the heat for steam reforming (an endothermic reaction), operate with an equivalence ratio in the range of 3 to 4, with a high level of steam (at least 30%) to increase the steam to carbon ratio (S/C1) to a value above 1 and dilute the concentration of O2. Addition of steam is necessary in such a system to prevent coke formation (carbon deposition) on the catalyst. However, addition of steam is not desirable because this water must be carried on the vehicle as a feed for the fuel processing system and the water would have to be very pure since typical impurities in tap water such as sodium, calcium, magnesium etc. are poisons for most reforming catalyst materials.